JP4856181B2 - Render a view from an image dataset - Google Patents

Abstract

The invention relates to a rendering system (100) for rendering a view from an image dataset, the rendering system comprising a selecting unit (110) for selecting a subset of the image dataset, a computing unit (120) for computing a first principal axis of a tensor on the basis of the subset of the image dataset, and a rendering unit (130) for rendering the view on the basis of the first principal axis. Using the information about the directionality and orientation of a structure, comprised in the selected subset of the image dataset and extracted from the first principal axis of the tensor, the rendering system (100) is arranged to effectively assist the user in selecting an advantageous view from the image dataset.

Description

  The present invention relates to a rendering system for rendering a view from an image data set.

  The invention further relates to an image acquisition system for acquiring an image data set comprising the rendering system.

  The invention further relates to a workstation comprising the above rendering system.

  The invention further relates to a rendering method for rendering a view from an image data set.

  The invention further relates to a computer program product loaded with a computer device having instructions for rendering a view from an image data set.

  An example of a method of the type described in the introductory part can be found in Jane P., published in Journal of Evaluation of Thoracic Imaging 19 (2004) 136-155. Ko and David P.M. It is known from the article “Computer Aided Diagnosis and the Evaluation of Lung Disease” by Naidich. The paper describes a computer program product for interactive lung nodule assessment. The method for displaying and navigating the 3D image data set used by the computer program product utilizes wheels that allow a 180 degree rotation of a relatively small VOI (Volume Of Interest) with lung nodules. The direction of the wheel determines the direction of the VOI. The VOI projection sequence acquired by changing the direction of the wheel can be displayed as a moving image. Viewing wheel projections helps to facilitate and facilitate manual navigation in 3D image datasets and to select the optimal view of nodules that are configured in the VOI. However, the method still requires the user to select the optimal view of the VOI. This is inconvenient and relatively time consuming and generates errors.

Furthermore, US 2005/0105829 describes a method for aligning tubular configurations in digital images. The method includes calculating a basic structure tensor of a selected point in the image volume and detecting eigenvalues and eigenvectors of the structure tensor based on the image gradient calculated in the vicinity of the selected point. In one implementation, the method consists of calculating a cartwheel projection about an axis defined by an eigenvector aligned with the tubular configuration. For this reason, the method assists the user in selecting the optimal view of the VOI. However, this method does not disclose a method for selecting the optimal neighborhood of the selected point.

  It is an object of the present invention to provide a rendering system of the type described in the introduction that is configured to effectively assist a user in selecting an effective view from an image data set.

The object of the present invention is a rendering system for rendering a view from an image data set, wherein a selection unit for selecting a subset of the image data set, and a first principal axis of a tensor based on the subset of the image data set. a calculation unit for calculations, and rendering the view on the basis of the first spindle, said to have a rendering unit for rendering a view from an image dataset, the selection unit, a plurality of candidate subsets of the image data set Solved by a rendering system configured to evaluate by calculating each feature suitable for the candidate subset and to select a subset of the image data set based on the evaluation .

  The selection unit is configured to select a subset of an image data set consisting of a configuration or a part thereof, hereinafter referred to as a configuration or structure. An example of such a configuration is a pulmonary nodule. A subset of the image data set is utilized by the computing unit to capture information about the orientation and orientation of the composition. This is accomplished by calculating the first principal axis of the tensor based on a subset of the image data set. The tensor is, for example, a structure tensor. The definition and properties of the structure tensor of a multidimensional image dataset are described in the paper “A tenor analysis for local structure in multi-dimensional imaging” by Horst Haubecker and Berne Jehne, referred to hereinafter as reference 1. Analysis and Synthesis '96, Ed.by B. Griet et al, Sankt Augustin 1996, 171-178). In order to qualify a composition, it is effective to render an optimized view from an image data set having that composition. The first principal axis of the tensor effectively defines an approximate principal axis configured in the image data set. The rendering unit is configured to render a view from the image data set based on the first major axis. The view can be a cross section of the image data set with planes substantially equal to each other in a plane orthogonal to the first principal axis of the structure tensor. Alternatively, the view can be a cross-section of the image data set by planes that are substantially equal to each other in the plane having the first principal axis of the structure tensor. The rendered view can be used by a healthcare professional such as a radiologist to qualify the composition contained in the image data set. User input for selecting a subset of the image data set has only manual exchanges such as the placement of the mouse pointer on the configuration, and optionally includes a mouse click thereafter. Thereby, the rendering system is configured to effectively assist the user in selecting an effective view from the image data set.

The usefulness of tensors calculated based on a subset of the image data set depends on the quality of the selection of the subset. If the subset is too small, the tensor may describe the local nature of the subject configuration. If the subset is too large, the tensors may be isotropically averaged in a number of configurations included in the subset of the image data set that results in a loss of meaningful direction information. A useful qualified feature is an indicator of direction. There are many indicators of directionality. Useful indicators of directionality can be extracted from the main component of the tensor. A useful feature that qualifies the candidate subset is the sum of the squares of the differences between the main components of the tensor. The selection unit is configured to select a subset corresponding to the maximum value of the sum.

  The rendering system of the present invention is useful for rendering views from multidimensional image data sets, particularly 3D and / or 4D image data sets. Image data sets include MRI (Magnetic Resonance Imaging), CT (Computed Tomography), Ultrasound, PET (Positron Emission Tomography), and SPECT (Simple Photo Emulation of any of the Single Photo Competitive Computations). .

  In an embodiment of the rendering system according to the invention, the view is a cross section of the image data set by planes substantially equal to each other in the plane having the first major axis. This view is particularly effective for elongated configurations such as blood vessels or parts thereof, hereinafter referred to as blood vessels, when the first major axis corresponds to the smallest major component of the tensor.

  In a further configuration of the rendering system according to the invention, the view is a projection of an image data set along an axis substantially equal to each other on the first main axis. This view is particularly effective for elongated configurations such as blood vessels when the first major axis corresponds to the largest principal component of the tensor. MIP (Maximum Intensity Projection) projection or IsP (Iso-surface Projection) projection can be used to render a view from an image data set that shows an elongated configuration with a principal axis orthogonal to the projection direction.

  In a further embodiment of the rendering system according to the invention, the calculation unit is arranged to calculate the second principal axis of the tensor and the view is an image with a plane substantially equal to a plane having a first principal axis and a second principal axis. A cross section of the data set. This view is particularly effective for performing measurements because it estimates the size of the displayed configuration.

  In a further embodiment of the rendering system according to the invention, the rendering system has a user interface for communicating with the rendering system. By combining the rendering system and the user interface, the user can communicate with the rendering system. The user interface is configured to display a view from the image data set to the user to select a target configuration. The user interface is further configured to assist in selecting a subset of the image data set. Optionally, the user interface selects how the user selects a subset of the image data set and / or selects a view type for rendering and accepts a selection method and / or a user selection of the view type. Configured to prompt.

  In a further embodiment of the rendering system according to the invention, the tensor is any one of a structure tensor, an inertia tensor and a Hessian matrix. Depending on the application and image dataset, the most effective of these tensors can be selected to implement the rendering system of the present invention. Optionally, the rendering system can be configured to select one of these tensors based on the evaluation and / or input.

  In a further embodiment of the rendering system according to the invention, the selection unit is configured to select a subset of the image data set based on a seed. The seed is entered utilizing a user interface configured to display a view from the image data set. The user can visually confirm the image data set to detect the target configuration. The user can then click the mouse at a position near the center of the composition. The selection unit is configured to accept this position. The location defines a seed for a subset of the image data set. The selection unit is further configured to expand the subset of the image data set based on the seed.

In a further embodiment of the rendering system according to the invention, the selection unit is configured to add the element to a subset of the image data set based on the distance between the element of the image data set and the seed. If the distance between the seed and the elements of the image data set is less than a predetermined value, the selection unit is configured to add a subset of the image data set .

In a further embodiment of the rendering system according to the invention, the selection unit is further configured to select a weighting set for weighting each element of the subset of the image data set, the calculation unit further comprising a first main axis and optionally Is configured to calculate other tensor characteristics based on a weighted set. This allows for weighting the contribution of each element from the selected subset to the first principal axis and other calculated tensor properties. The weight of each element can be based on, for example, the distance between that element and the seed. One skilled in the art will recognize that this general approach is equivalent to replacing a subset of the image data set with a window function, as described in Chapter 3 of Reference 1.

A further subject of the present invention is an image acquisition system of the type as described in the introductory paragraph, which is configured to effectively assist the user in selecting an effective view from the image data set. A rendering system for rendering a view from an image dataset, a selection unit for selecting a subset of the image dataset, and a calculation unit for calculating a first principal axis of a tensor based on the subset of the image dataset When the first rendering the view on the basis of the spindle, have a rendering unit for rendering the view from the image dataset, the selection unit, a plurality of candidate subsets of the image data set, the candidate subsets Solved by an image acquisition system having a rendering system configured to evaluate by calculating each feature suitable for and selecting a subset of the image data set based on the evaluation .

A further object of the present invention is a workstation of the type as described in the introductory paragraph, which is configured to effectively assist the user in selecting an effective view from the image data set. A rendering system for rendering a view from an image dataset, a selection unit for selecting a subset of the image dataset, and a calculation unit for calculating a first principal axis of a tensor based on the subset of the image dataset When the first rendering the view on the basis of the spindle, have a rendering unit for rendering the view from the image dataset, the selection unit, a plurality of candidate subsets of the image data set, the candidate subsets Solved by a workstation having a rendering system configured to evaluate by calculating each feature suitable for and selecting a subset of the image data set based on the evaluation .

  A further subject of the present invention is a rendering method of the type as described in the introductory paragraph, which is configured to effectively assist the user in selecting an effective view from the image data set. This is a rendering method for rendering a view from an image data set, the selection step selecting a subset of the image data set, and the calculating step calculating a first principal axis of a tensor based on the subset of the image data set; Rendering the view based on the first principal axis and rendering the view from the image data set.

  A further subject of the present invention is a computer program of the type as described in the introductory paragraph, which is configured to effectively assist the user in selecting an effective view from the image data set. This is a computer program loaded by a computer device, said computer program having instructions for rendering a view from an image data set, said computer device comprising a processing unit and a memory, The computer program, after being loaded, selects a subset of the image data set, a calculation step of calculating a first principal axis of a tensor based on the subset of the image data set, and based on the first principal axis Rendering a view and rendering a view from the image data set are solved by a computer program that causes the processing unit to perform.

  Modifications and variations of the image acquisition system, workstation, rendering method and / or computer program corresponding to the described rendering system improvements and variations can be performed by those skilled in the art based on this description.

  These and other features of the rendering system, image acquisition system, workstation, rendering method and computer program according to the present invention will be apparent with respect to implementations and embodiments described below with reference to the accompanying drawings.

FIG. 1 schematically illustrates an embodiment of a rendering system 100 that renders views from an image data set. The rendering system
A selection unit 110 for selecting a subset of the image data set;
A calculation unit 120 for calculating the first principal axis of the tensor based on a subset of the image data set;
A rendering unit 130 for rendering a view based on the first principal axis, thereby rendering a view from the image data set;
Have

Optionally, the rendering system 100
A user interface 140 for communicating with the rendering system 100;
A memory unit 170 for storing input / output data;
Have

  In the embodiment of the rendering system 100 shown in FIG. 1, there are two input connectors 181, 182 and 183 for input data. The first input connector 181 is configured to receive data input from a data storage such as a hard disk, a magnetic tape, a flash memory, or an optical disk. The second input connector 182 is configured to receive data input from a user input device such as a mouse or a touch screen. The third input connector 183 is configured to receive data input from a user input device such as a keyboard. The input connectors 181, 182 and 183 are connected to the input control unit 180.

  In the embodiment of the rendering system 100 shown in FIG. 1, there are two output connectors 191 and 192 for output data. The first output connector 191 is configured to output data to a data storage such as a hard disk, a magnetic tape, a flash memory, or an optical disk. The second output connector 192 is configured to output data to the display device. The output connectors 191 and 192 receive each data via the output control unit 190.

  Those skilled in the art will appreciate that there are many ways to connect the input connectors 181, 182 and 183 of the rendering system 100 to the input device, and to connect the output connectors 191 and 192 of the rendering system 100 to the output device. Let's go. These methods include, but are not limited to, wired and wireless connections, digital networks such as local area networks (LAN) and wide area networks (WAN), the Internet, digital telephone networks and analog telephones. Consists of a net.

  In an embodiment of the rendering system 100 according to the present invention, the rendering system has a memory unit 170. The memory unit 170 is configured to receive input data from an external device via any of the input connectors 181, 182, and 183 and store the received input data in the memory unit 170. Loading the data into the memory unit 170 allows for quick access to the data portion associated with each unit of the rendering system 100. The input data is composed of an image data set. The memory unit 170 can be realized by a device such as a RAM (Random Access Memory) chip, a ROM (Read Only Memory) chip, and / or a hard disk. Preferably, the memory unit 170 has a RAM for storing the image data set. The memory unit 170 is also configured to receive data from and provide data to each unit of the rendering system 100 having the selection unit 110, the calculation unit 120, and the rendering unit 130 via the memory bus 175. The memory unit 170 is further configured to make the data available to an external device via any of the output connectors 191 and 192. The storage of data from each unit of rendering system 100 in memory unit 170 effectively improves the performance of each unit of rendering system 100 and transfers data from each unit of rendering system 100 to an external device. Improve the rate.

  Alternatively, the rendering system 100 does not have the memory unit 170 and the memory bus 175. Input data used by the rendering system 100 is provided by at least one external device such as an external memory or a processor connected to each unit of the rendering system 100. Similarly, output data generated by the rendering system 100 is provided to at least one external device such as an external memory or processor connected to each unit of the rendering system 100. The units of the rendering system 100 are configured to receive each other via an internal connection or a data bus.

  In a further embodiment of the rendering system 100 according to the present invention, the rendering system has a user interface 140 for communicating with the rendering system 100. The user interface 140 includes a display unit that displays data to the user and a selection unit that performs selection. By combining the rendering system 100 and the user interface 140, the user can communicate with the rendering system 100. The user interface 140 is configured to display a view from the image set to the user for selecting a target configuration. The user interface 140 is further configured to assist in selecting a subset of the image data set. Optionally, the user interface may have multiple processing modes of the rendering system, such as a mode that determines how to select a subset of the image dataset and the type and / or number of views rendered from the image dataset. Is possible.

  Alternatively, the rendering system 100 can utilize external input devices and / or external displays that are connected to the rendering system via input connectors 182 and / or 183 and output connectors 192. Those skilled in the art will appreciate that there are many user interfaces that can be effectively implemented as units of the rendering system 100 of the present invention.

  The selection unit 110 is configured to select a subset of the image data set. A subset of the image dataset is used to determine an effective view of the composition of the image dataset. Thus, the calculation unit 120 is configured to utilize a subset of the image data set to capture information regarding the directionality and orientation of the configuration. In an embodiment of the rendering system 100 according to the present invention, a view, such as an axial slice rendered from an image data set, is displayed on a display device. The user uses the mouse to position the mouse pointer and click the mouse button on the configuration to select the configuration that the user desires to view. The selection unit 110 is configured to receive a position selected by the user and use the position to select an image element in the vicinity of the position, preferably the closest element, as a seed element, hereinafter referred to as a seed. The Seeds are further used to augment a subset of the image data set. Each element from the image data set is included with a subset of the image data set such that the distance between the element and the seed element is less than a predetermined distance threshold. The distance is a geometric distance such as a Euclidean distance between the seed position and the element position. Alternatively, the distance can be a phase distance.

  In a further embodiment of the rendering system 100 of the present invention, the seed has a plurality of elements from the image data set. Using the mouse, the user can surround the area of the displayed view. The seed has multiple elements from the image data set in the enclosed area. Each element from the image data set is then added to the selected subset such that the distance between the element and the element from the seed is less than a predetermined distance threshold.

  In a further embodiment of the rendering system 100 according to the present invention, the user interface 140 for selecting a subset of the image data set includes a plurality of containers, such as spheres. The user selects a container from the menu and places the selected container in the displayed view of the image data set. A subset of the selected image data set has each element in the container. Optionally, the user interface may allow shaping of a sphere into an ellipsoid, shaping an ellipsoid, transforming a sphere or ellipsoid, and rotating the ellipsoid.

  In a further embodiment of the rendering system 100 according to the invention, the selection unit 110 is configured to allow selection of a subset of the image data set based on image segmentation. The user selects a seed that is included in the configuration identified during image segmentation. This has the effect that a subset of the selected image data set has all elements belonging to the selected configuration. The result of the image segmentation is stored in the memory unit 170. Alternatively, the selection unit 110 has a segmentation system that depicts a configuration with a seed. A segmentation system that is fast enough for the rendering system is available for this purpose.

  A calculation unit 120 is configured to calculate a first principal axis of the tensor based on a subset of the image data set. Optionally, the calculation unit is also configured to calculate other properties of the tensor. A tensor characteristic may, for example, consist of each component of the tensor and other principal axes. In an embodiment of the rendering system 100 according to the present invention, a plurality of characteristics are comprised of each component of the structure tensor, the principal component of the structure tensor, and the principal axis of the structure tensor in the coordinate system of the image data set. The structure tensor J (P) at position P is the tensor product of the gradient vector G (P) at P and the gradient vector G (P) at P. For this reason, the tensor J (P) is a quadratic function of the gradient vector G (P). Therefore, J (P) is independent from the sign of the gradient vector G (P). However, the structural tensor still has information about the orientation and orientation of the configuration at position P. Furthermore, the structure tensor J (P) is positive definite so that all main components of the structure tensor are non-negative. Section 3 of Reference 1 and A.I. Ravishankar Rao and Brian G. In “Computing orientated texture fields” by Schnuck (Computer Vision and Pattern Recognition, CVPR '89, IEEE Computer Society, June 4-8, 1989, pages 61-68) is described as a set. Yes. The structure tensor describes the directionality and orientation of the configuration at position P. The directionality and orientation of the structure tensor J (P) in the neighborhood U of P with the position of each element from a subset of the image data set is described by the sum of the structure tensors J (P ′) at all positions P ′ of U. Is done. This sum is also referred to as the structure tensor J (P). The position P can be any position in U. The position P defines the origin of the coordinate system with respect to the components of the structural tensor and the principal axis. In the following, this position is referred to as the structure position. The usual choice of structure location is the seed point. Alternatively, the structure position P is the geometric center of U or the center of mass of a subset of the image data set. The neighborhood U is uniquely equal to a subset of the image data set. U is equal to the projection of a subset of the image data set to the spatial domain or spatio-temporal domain. A subset of the image data set is a set of pairs (P ′, I (P ′)) for all P ′ belonging to U. Here, I (P ′) is the intensity at the position P ′.

  Section 3 of reference 1 describes a method for calculating structural tensor components from an image data set. The calculated structural tensor consists of 10 4D independent components, 6 3D independent components, or 3 2D independent components. The calculated tensor component is stored in the memory unit 170 for further processing by the calculation unit 102 and / or use by the rendering unit 130. Although the description of the present invention is based on 3D images, the present invention is applicable to any multidimensional image data set.

There are multiple methods for calculating the algebraic and numerical tensor principal components {T _ξ , T _η , T _ζ } and principal axes {ξ, η, ζ}, for example, Lloyd N. Chapter 5 and William H. of “Numerical Linear Algebra” (Soc for Industrial & Applied Math, May 1, 1997) by Trefethen and David Bau. Press, Brian P.M. Flannery, Saul A. et al. Teukolsky and William T. “Numerical Recipes in C: The Art of Scientific Computing” by Vetterling (Cambridge University Press; 2nd edition, October 30 and 1992). The six spherical components {T ₀₀ , T _2-2 , T _2-1 , T ₂₀ , T ₂₁ , T ₂₂ } of the symmetric tensor are converted into six Cartesian components {T _xx , T _xy , T _xz , T _yy , T _yz , T _zz }.

ここで、ｉは虚数単位を表す。効果的には、球体コンポーネントはまた主コンポーネント｛Ｔ_ξ，Ｔ_η，Ｔ_ζ｝から求めることも可能である。

Here, i represents an imaginary unit. Effectively, the spherical component can also be determined from the main components { _Tξ , _Tη , _Tζ }.

等方性構成について、テンソルのすべての主コンポーネントは同一であり、このため、Ｔ_２０とＴ_２±２（すなわち、Ｔ_２２とＴ_２−２）はゼロとなる。コンポーネントＴ_００はまた、テンソルの等方性コンポーネントとして参照される。ときどき、Ｔ_００の代わりに、非正規化等方性コンポーネントが使用される。

For isotropic configurations, all major components of the tensor are identical, so T ₂₀ and T _{2 ± 2} (ie, T ₂₂ and T _2-2 ) are zero. Component _T00 is also referred to as an isotropic component of the tensor. Sometimes, instead of T _00, denormalized isotropic component is used.

Ｔ_ｉｓｏはテンソルの任意のデカルト座標セットに対して同じであるため、それはまた、

Since T _iso is the same for any Cartesian coordinate set of tensors, it also

から求めることも可能である。画像データセットが一様である場合、Ｔ_ｉｓｏ＝０となる。

It is also possible to obtain from. If the image data set is uniform, T _iso = 0.

For configurations with axial symmetry having a ζ principal axis selected to be a rotational axis of a configuration such as tubular or cylindrical, component T _{2 ± 2} (ie, T ₂₂ and T _2-2 ) Becomes zero. This means that T _ξ and T _η are equal to each other. Component _{T 20} quantifies the deviation (deviation) of the structure from an isotropic symmetry. The lower the isotropic this configuration, the absolute value of T ₂₀ is increased. _Sometimes, instead of _{T 20,} denormalized component

が利用される。

Is used.

The last two components, T _{2 ± 2} (ie, T ₂₂ and T _2-2 ) are equal to each other. Assuming that the main rotational axis is selected to be the ζ main axis, these two components quantify how much the symmetry of the configuration deviates from the axial symmetry. The lower the axial symmetry of the configuration, the greater the absolute value of T _{2 ± 2} . Sometimes denormalized components instead of T _{2 ± 2}

が利用される。

Is used.

  There are numerous embodiments of the rendering system 100 according to the present invention that differ in the properties of the tensors calculated by the calculation unit 120. The described embodiments are illustrative of the invention and are not intended to limit the scope of the claims. For example, the calculation unit 120 can be configured to calculate six Cartesian tensor components according to the coordinate system of the image data set. The calculation unit 120 can be further configured to determine the smallest principal component of the tensor and the principal axis corresponding to the smallest principal component using the Ritz variational method. This example is particularly suitable for rendering a view of an elongated configuration, such as a blood vessel, whose major axis corresponding to the smallest major component of the tensor is substantially equal to the major axis of the configuration near the seed. Alternatively, the calculation unit 120 can be configured to calculate six Cartesian tensor components in the coordinate system of the image data set. The calculation unit 120 can be further configured to determine a principal axis corresponding to the largest principal component largest principal component of the tensor using a Ritz variational method. This embodiment is also particularly effective in rendering a view of an elongated configuration, such as a blood vessel, whose major axis corresponding to the largest principal component of the tensor is substantially equal to an axis orthogonal to the major axis of the configuration.

  The rendering unit 130 is configured to render a view from the image data set based on the first principal axis and the configured position. In an embodiment of the rendering system 100 according to the present invention, the rendering unit 130 renders three cross-sectional views from the image data set. The calculation unit 120 is configured to calculate the three principal axes of the tensor. Each pair of principal axes defines a viewing plane. The rendering unit 130 is configured to render three cross-sectional views of the image data set with three planes, each plane being substantially equal to each other in the planes constituting the principal axis pair.

  In an embodiment of the rendering system 100 according to the present invention, the rendering unit 130 renders a cross-sectional view from the image data set. The view corresponds to a cross section of the image data set by planes substantially equal to each other in the plane having the first principal axis of the tensor calculated by the calculation unit. This embodiment is particularly effective for rendering a view of an elongated configuration, such as a blood vessel, when the first major axis corresponds to the smallest major component of the tensor. In this case, the first main axis defines directions that are substantially equal to the main axis of the configuration. Since the blood vessel has a substantially axial local symmetry having a main axis which is an axis having local axial symmetry, planes substantially equal to one of the planes having the main axis are suitable for display. The rendering unit 130 is configured to select one display surface. Optionally, rendering unit 130 and user interface 140 allow the user to select another display surface or rotate the display surface about an axis substantially equal to the first major axis and can be displayed as a movie. Can be configured to allow rendering of a random view sequence. Alternatively, if the second major axis is calculated or known, the view may be a planar cross-sectional view defined by planes that are substantially equal to a plane having the first major axis and the second major axis. This configuration is particularly suitable for measuring the distance between each element of the structure and the size of the structure.

  In an embodiment of rendering system 100 according to the present invention, rendering unit 130 renders a 3D view from an image data set. The view is a projection of the image data set along an axis that is substantially equal to the first major axis. The present invention is not limited to any particular image rendering technique. For example, a MIP (Maximum Intensity Projection) algorithm or an IsP (Iso-surface Projection) algorithm can be used. In MIP, the pixel is set to the maximum value along the ray. In IsP, rays are terminated in an iso-surface. An isosurface is defined by the positions of elements having the same value. Further information on image rendering can be found in “Introduction to Volume Rendering (Hewlett-Packard Professional Books)” (Princess Hall 98) by Barthold Richtenbelt, Randy Crane and Shaz Naqvi. Projection views are particularly effective for elongated configurations such as blood vessels when the first principal axis corresponds to the largest principal component of the tensor. Such a first main axis is perpendicular to the main axis corresponding to the smallest main component substantially equal to the main axis of the configuration. Thus, the projection effectively displays an elongated configuration having a main axis that is substantially perpendicular to the projection direction. If the view of the configuration covered by other configurations is unclear, the configuration that makes the view unclear can be removed, or its transparency can be seen in the paper “Importance by Ivan Viola, Armin Kanitsar and Meister Edward Groeller. can be increased as described in "drive volume rendering" (IEEE Visualization 2004, October 10-15, Austin, Texas, USA).

  Optionally, the user interface 140 can be configured to allow the user to select a view to render from the image data set. The rendering system 100 can be configured to calculate the properties of the tensor for rendering the selected view and render the view. Optionally, each view can be scaled, such as magnified and / or clipped.

  There are many ways for a person skilled in the art to select a subset of the image dataset, select useful properties of the tensor computed by the calculation unit 120, and render a view from the image dataset based on the selected useful properties. You will recognize that. The described embodiments are illustrative of the invention and are not intended to limit the scope of the claims.

  There are several tensors that can generate information about the orientation and orientation of the composition of the image dataset. In an embodiment of the rendering system 100 according to the present invention, the tensor is a structural tensor. The structure tensor is an important indicator of the directionality and orientation of the image data set. Alternatively, inertia tensors can be used to define the orientation and orientation of the composition in the image data set. Inertial tensors are particularly useful when applied to objects in segmented images. A further option is to use the second order partial derivative Hessian of the image data set. The eigenvalues and eigenvectors of the Hessian can be used in place of the main component and main axis of the tensor. Optionally, the calculation unit 120 may calculate other tensors and / or other matrix properties of each tensor and / or matrix from any plurality of tensors and / or matrices, along with the first principal axis and / or the first eigenvector. It can be configured to calculate. The calculation unit 120 further combines the usefulness of other arbitrarily calculated tensor and / or matrix properties along with the calculated first principal axis and / or first eigenvector to render a view from the image data set. It can be configured to evaluate. For example, the computing unit can be configured to select a tensor or matrix having a maximum anisotropy defined by the difference between the largest principal component or eigenvector and the smallest principal component or eigenvector. Alternatively, the user may select a tensor or matrix that is used to render the view from the image data set.

  Those skilled in the art will determine the orientation and orientation of the composition in the image dataset as disclosed in the present invention, and based on the first principal axis and / or the first eigenvector of the tensor or matrix, It will be appreciated that there are other tensors and / or matrices that can be used to select a view. The structural tensors, inertia tensors and Hessian are used to explain the effects of the present invention and do not limit the scope of the claims. Throughout the description of the present invention, “tensor”, “principal component” and “principal axis” mean “matrix”, “eigenvalue” and “eigenvector”, respectively.

  As already emphasized, proper selection of a subset of the image data set is essential to better determine the first principal axis and optionally other tensor characteristics. Thus, in an embodiment of the rendering system 100 according to the present invention, the selection unit 110 evaluates a plurality of candidate subsets of the image data set by calculating each feature of the plurality of candidate subsets, and based on the evaluation, Configured to select a subset. The user selects a seed and the selection unit 110 is configured to select a subset of the image data set having elements from the image data set such that the distance between the seed and the element is less than a predetermined minimum distance. Is done. Next, the selection unit 110 is configured to calculate features for limiting a subset of the image data set. Preferably, the selection unit 110 is configured to use the calculation unit 120 to perform the calculation. There are many indicators of directionality. Some useful direction indicators can be extracted from the main components of the tensor.

  The normalized direction indicator is

の式により与えられる。効果的には、Ｄ_１は球体コンポーネントセットから、主コンポーネントセットから、及びデカルトコンポーネントセットから求めることができる。等方性テンソルについて、方向性の指標の値はゼロとなる。１つの小さな主コンポーネントと２つのやや大きな主コンポーネントを有する軸方向に対称的なテンソルに対して、方向性の指標は約０．５となる。この値又は類似する値（すなわち、０．５に近接した）は、血管など、回転軸にパラレルな方向に相対的に一様な強度を有する長い管状又は円筒状構成に対して想定できる。２つの小さな主コンポーネントと１つの相対的に大きな主コンポーネントを有する軸方向に対称的なテンソルについては、方向性の指標は約２となる。この値又は類似する値（すなわち、２に近接した）は、回転軸に直交した各平面において相対的に一様であって、回転軸の方向に急激に変化する強度、すなわち、強力な勾配を有する細長い軸方向に対称的な構成に対して想定できる。

Is given by Effectively, D ₁ can be determined from the spherical component set, from the main component set, and from the Cartesian component set. For isotropic tensors, the value of the directionality index is zero. For an axially symmetric tensor with one small main component and two slightly larger main components, the directionality index is about 0.5. This value or a similar value (ie, close to 0.5) can be envisioned for a long tubular or cylindrical configuration having a relatively uniform strength in a direction parallel to the axis of rotation, such as a blood vessel. For an axially symmetric tensor with two small main components and one relatively large main component, the directionality index is about 2. This value or a similar value (ie, close to 2) is relatively uniform in each plane perpendicular to the axis of rotation, and gives an intensity that changes rapidly in the direction of the axis of rotation, ie, a strong gradient. It can be envisaged for an elongated axially symmetric configuration.

  Alternatively, for a cylindrical configuration having cylindrical axes that are substantially equal to the ζ main axis, such as a blood vessel, a useful directional index is given by:

または、主コンポーネントＴ_ξ≧Ｔ_η≧Ｔ_ζ≧０について、

Or for the main component T _ξ ≧ T _η ≧ T _ζ ≧ 0,

を利用することが可能である。

Can be used.

  One skilled in the art will appreciate that other directional indicators exist and that the present invention is not limited to any particular choice of directional indicators.

  After calculating the directional index for a predetermined minimum radius, the selection unit 110 is configured to repeatedly increment the radius by a predetermined increment and recalculate the directional index for each radius to a predetermined maximum radius. Each value of the radius and direction index is stored in the memory unit 170 for evaluation. The optimal value for the predetermined minimum radius and the predetermined maximum radius, and the optimal value for the predetermined increment, depending on the image data set and the configuration being viewed, may be parameters defined by the user. When the value of the directional index is calculated for each candidate subset of the image data set, the selection unit 110 may select the optimal subset of the image data set that corresponds to the optimal value of the directional index, such as the maximum value of the directional index. Configured to select. Alternatively, the selection unit 110 can utilize other algorithms for observing the optimal value of the feature, such as a steepest ascent algorithm. The optimal subset is then utilized by the rendering system 100 to render the view from the image data set.

  In a further embodiment of the rendering system 100 according to the present invention, the selection unit 110 is further configured to select a weight set for weighting each element of the subset of the image data set, and the calculation unit 120 is further based on the weight set. The first principal axis and optionally other tensor characteristics are calculated. This allows weighting of the contribution of each element from the selected subset to the first principal axis and other calculated tensor properties. The weight of each element can be based on, for example, the distance between that element and the seed. Alternatively, weights can be assigned by the user. One skilled in the art will recognize that this general approach is equivalent to replacing a subset of the image data set with a window function, as described in Chapter 3 of Reference 1.

  Those skilled in the art will appreciate that other embodiments and implementations of the rendering system 100 according to the present invention are also possible. In particular, it is possible to redefine each unit of the system and reassign their functions. For example, in an embodiment of the rendering system 100 of the present invention, the calculation unit 120 can be divided into a selection unit 110 and a rendering unit 130. The portion configured in the selection unit 110 can be configured to perform the calculations necessary to evaluate the candidate image data set, and the portion configured in the rendering unit 130 renders a view from the image data set. It can be configured to calculate the first principal axis and other features of the tensor needed for In a further embodiment of the rendering system 100 according to the invention, the calculation unit 120 of the embodiment described above and a plurality of calculation units can be replaced, each calculation unit having a first principal axis of a different tensor and optionally other characteristics. And is configured to calculate The user interface 140 can be configured to prompt the user to select a first principal axis to render a view from the image data set based on an evaluation of the tensor used.

  The selection unit 110, the calculation unit 120, and the rendering unit 130 can be implemented using a processor. Usually, these functions are performed under the control of a software program product. The software program product is running. It is loaded into a memory such as a RAM and executed from there. The program may be loaded from ROM, hard disk or background memory such as magnetic and / or optical storage, or may be loaded via a network such as the Internet. Optionally, an application specific integrated circuit may provide the described functionality.

  There are many possible applications of the rendering system 100 of the present invention. An effective application is the application of the rendering system 100 to a medical image data set. The rendering system 100 of the present invention can also be useful in other applications. For example, the rendering system 100 may be useful in cell form to render an effective view of the cell configuration.

  FIG. 2 shows a cross section in the axial slice 210 and three effective cross sections 220, 230 and 240. An axial slice 210 of the patient's chest is shown on the left. The arrow indicates a white dot in the image. This may be a nodule or blood vessel cross section. Two possibilities cannot be determined based on the axial slice. To obtain a better view of the composition represented by the dots in the left image, the user can select a seed for a subset of the image data set by moving the mouse pointer and clicking a mouse button near the dot. Thereafter, the selection unit 110 is triggered to select the optimal subset of the image data set, and the calculation unit 120 is configured to calculate the three principal axes and the structure position of the structure tensor. The rendering unit renders the three cross-sections 220, 230 and 240 of the configuration with three planes defined by three pairs of major axes. Three magnified and clipped views 220, 230, and 240 of the configuration are shown on the right side of the axial slice 210. In the cross-sectional view 220, it can be clearly confirmed that the configuration is not a nodule but a blood vessel.

  FIG. 3 schematically illustrates an embodiment of an image acquisition system 300 that utilizes the rendering system 100 of the present invention. The image acquisition system 300 includes an image acquisition system unit 310 connected to the rendering system 100 via an internal connection, an input connector 301, and an output connector 302. This configuration effectively improves the function of the image acquisition system 300 that provides the image acquisition system 300 with the effective image browsing function of the rendering system 100. This viewing function is particularly useful when the image acquisition system 300 is further configured for interactive image acquisition, and thus allows the operator to determine the data to be acquired based on the displayed image. might exist. Specific examples of the image acquisition system include a CT system, an X-ray system, an MRI system, an ultrasound system, a PET (Positron Emission Tomography) system, and a SPECT (Single Photon Emission Computed Tomography) system.

  FIG. 4 schematically illustrates an example of a workstation 400. The workstation has a system bus 401. A processor 410, a memory 420, a disk input / output (I / O) adapter 430 and a user interface (UI) 440 are operatively connected to the system bus 401. The disk storage device 431 is operatively connected to the disk I / O adapter 430. A keyboard 441, mouse 442 and display 443 are operatively connected to the UI 440. The rendering system 100 of the present invention is realized as a computer program and stored in a disk storage device 431. The workstation 400 is configured to load a program and input data into the memory 420 and execute the program on the processor 410. A user can enter information into workstation 400 using keyboard 441 and / or mouse 442. The workstation is configured to output information to the display device 443 and / or the disk 431. Those skilled in the art have numerous other workstation embodiments known in the art, which are intended to illustrate the invention and limit the invention to those embodiments. Should not be interpreted.

  FIG. 5 schematically illustrates an example of a rendering method 500. The START step 501 begins rendering a view from the image data set according to the method 500. In a selection step 510, a subset of the image data set is selected. The subset is composed of elements of an image data set in a certain sphere. The center of the sphere is at the seed location selected by the user. The radius of the sphere is also determined by the user. In a calculation step 520, the three principal axes of the structure tensor are calculated based on a subset of the image data set. In a rendering step 530, a view from the image data set is rendered based on the three principal axes of the structure tensor. Each view is a cross-sectional view of the image data set with planes substantially equal to each other in planes having different principal axis pairs of the structural tensor. The END step 599 ends the rendering of the view from the image data set by the method 500.

  The order of the described embodiments of the method of the present invention is not compulsory, and those skilled in the art will utilize a thread processing model, multiprocessor system or multiprocess without departing from the idea as intended by the present invention. Thus, the order of the steps may be changed, or the steps may be executed simultaneously.

  The rendering system 100 of the present invention may be implemented as a computer program product and can be stored on any suitable medium such as a magnetic tape, magnetic disk, or optical disk. This computer program can be loaded into a computer configuration having a processing unit and a memory. Once loaded, the computer program product provides the processing unit with the ability to perform rendering tasks.

  The above-described embodiments are intended to illustrate rather than limit the invention, and it will be appreciated by those skilled in the art that other embodiments can be designed without departing from the scope of the appended claims. Should be noted. In the claims, any reference signs placed between parentheses shall not be construed as limiting the claim. The term “comprising” does not exclude the presence of elements or steps not listed in a claim. The word “a” preceding an element does not exclude the presence of a plurality of such elements. The present invention can be realized by hardware having a plurality of elements and an appropriately programmed computer. In a system claim listing multiple units, some of these units can be implemented by the same hardware or software item. The use of words such as “first”, “second” and “third” does not indicate any ordering. These words are interpreted as names.

FIG. 1 schematically illustrates an embodiment of a rendering system. FIG. 2 shows an axial slice cross section (left side) and three effective cross sections. FIG. 3 schematically shows an embodiment of the image acquisition system. FIG. 4 schematically shows an embodiment of a workstation. FIG. 5 schematically shows an embodiment of the rendering method.

Claims (13)

A rendering system that renders a view from an image dataset,
A selection unit for selecting a subset of the image data set;
A calculation unit for calculating a first principal axis of a tensor based on a subset of the image data set;
A rendering unit for rendering the view based on the first principal axis and rendering a view from the image data set;
Have
The rendering unit is configured to evaluate a plurality of candidate subsets of the image dataset by calculating each feature suitable for the candidate subset, and to select a subset of the image dataset based on the evaluation system.   The rendering system of claim 1, wherein the view is a cross section of the image data set with planes substantially equal to each other in a plane having the first major axis.   The rendering system of claim 1, wherein the view is a projection of the image data set along axes that are substantially equal to the first principal axis. The calculation unit is configured to calculate a second principal axis of the tensor;
The rendering system of claim 2, wherein the view is a cross-section of the image data set by planes that are substantially equal to each other in a plane having the first major axis and the second major axis.   The rendering system of claim 1, further comprising a user interface for communicating with the rendering system.   The rendering system according to claim 1, wherein the tensor is one of a structure tensor, an inertia tensor, and a Hessian matrix.   The rendering system of claim 1, wherein the selection unit is configured to select a subset of the image data set based on a seed.   The rendering of claim 1, wherein the selection unit is further configured to add elements of the image data set to a subset of the image data set based on a distance between the seed and elements of the image data set. system. The selection unit is further configured to select a weighted set for weighting each element of the subset of the image data set;
The rendering system of claim 1, wherein the calculation unit is further configured to calculate the first principal axis based on the weight set.   An image acquisition system for acquiring an image data set comprising the rendering system according to claim 1.   A workstation comprising the rendering system according to claim 1. A rendering method executed on a processor device for rendering a view from an image dataset,
A selection step for selecting a subset of the image dataset;
Calculating a first principal axis of a tensor based on a subset of the image data set;
Rendering the view based on the first principal axis and rendering a view from the image data set;
Have
The selecting step is configured to evaluate a plurality of candidate subsets of the image data set by calculating each feature suitable for the candidate subset, and to select the subset of the image data set based on the evaluation . A computer program loaded by a computer device, the computer program comprising instructions for rendering a view from an image data set, the computer device comprising a processing unit and a memory,
After the computer program is loaded,
A selection step of selecting a subset of the image data set;
Calculating a first principal axis of a tensor based on a subset of the image data set;
Rendering the view based on the first principal axis and rendering a view from the image data set;
To the processing unit,
The selecting step is configured to evaluate a plurality of candidate subsets of the image data set by calculating respective features suitable for the candidate subset, and to select the subset of the image data set based on the evaluation program.

Images